In HIV-infection, viral specific CTL are likely to play a crucial role in host defense. The principal investigator's preliminary analysis of CTL recognition of relatively conserved regions of gp160 in a genetically diverse population suggests a bias in MHC restriction elements used for the response, with a preferential usage of certain MHC alleles and a striking absence of epitopes restricted by the most prevalent MHC alleles which are expressed by 76% of the study population. One aim of this proposal is to extend these studies to include isolate-specific env epitopes and other HIV proteins to see if there is a similar bias in restriction elements used in their recognition. For these studies, peptide-specific T cell lines generated from stored samples of HLA typed seropositive subjects will be used to determine the restricting elements for immunodominant env, gag RT epitopes. Recent technical advances suggest that by experimental manipulation, it may be possible to diversify the immune response or generate a more effective immune response than is seen in natural infection. The second question this proposal addresses is whether it is possible to generate CTL response to particular epitopes even though they are not immunodominant in natural infection. Different strategies like using highly purified peptide-pulsed dendritic cells, modifying the APC to increase the binding of exogenously added peptides and over expression of peptides in APC in the form of minigenes will be tried to elicit responses to cryptic and subdominant epitopes. Peptides selected for high affinity binding to prevalent MHC as well as those seen as immunodominant in some HIV seropositive individuals will be used for the study. However, since in vivo effectiveness of CTL will ultimately depend on their ability to kill infected CD4+ cells, CTL directed against immunodominant epitopes or those induced experimentally will be tested for effectiveness in killing HIV infected CD4 targets. For these studies CD4+ cells uniformly expressing the virus will be generated using the infectious molecular clone of HIV, R7, neo containing a neomycin resistance marker. The virus infected cells selected in G418 will be used as targets in killing assays. These studies are likely to provide insights into the nature of the CTL response to HIV in an outbred population and may have implications for vaccine design and immunotherapy.